|
APIs in Category: dp-restore |
API version 3.6R2 |
dp-restore-to-new-path |
These operations are used by Freya to restore data from a backed-up data set. |
| dp-restore-to-new-path | [top] |
Start a restore operation on part of a data set. This operation copies whole members or its sub-paths of the data set from a specific backup version to a new location. The operation is performed in the background by a job.
Error conditions:
- EOBJECTNOTFOUND - The data set name or ID or one of the member name or ID was not found.
- EACCESSDENIED - The data set exists, but the user invoking the API has no DFM.BackupManager.Restore permission on the data set.
- EINTERNALERROR - An error occurred while processing the request. Try again later.
- EINVALIDINPUT - The number of paths to restore was either 0 or more than 1000 or neither backup-id nor backup-version was specified.
- EINVALIDHOST - One of the specified destination hosts was not a valid restore destination.
- EDATABASEERROR - A database error occured.
Input Name Range Type Description backup-id integer
optional
Identifier of the backup to restore from. If this parameter is specified, then backup-version is ignored.
Range: [1..2^31-1]backup-version dp-timestamp
optional
Timestamp of the backup which should be restored from. If the backup-id is not specified, this parameter is required. A backup location that matches this version will be picked to do the restore from. dataset-name-or-id obj-name-or-id
Name or identifier of the data set to restore part of. members dp-recover-member-info[]
List of members and their path to recover within the data set. Each member includes the new location to restore the data. The maximum number of members is 1000. migrate-relationship boolean
validate
optional
Flag indicating whether to migrate the relationships. By default, the value is false. Output Name Range Type Description job-id integer
Id of the job that handles the restore operation.
Range: [1..2^31-1]
Errno Description EOBJECTNOTFOUND EACCESSDENIED EINTERNALERROR EINVALIDINPUT EINVALIDHOST EDATABASEERROR
| dp-restore-to-primary | [top] |
Start a restore operation on part of a data set. This operation copies files and/or directories from a specific backup version back to the primary location. The operation is performed in the background by a job.
Error conditions:
- EOBJECTNOTFOUND - The data set name or ID or one of the member name or ID was not found.
- EACCESSDENIED - The data set exists, but the user invoking the API has no DFM.BackupManager.Restore permission on the data set or on any of the members.
- EINTERNALERROR - An error occurred while processing the request. Try again later.
- EINVALIDINPUT - The number of paths to restore was either 0 or more than 1000 or neither backup-id nor backup-version was specified.
- EDATABASEERROR - A database error occured.
Input Name Range Type Description backup-id integer
optional
validate
Identifier of the backup to restore from. If this parameter is specified, then backup-version is ignored.
Range: [1..2^31-1]backup-version dp-timestamp
optional
Timestamp of the backup which should be restored from. If the backup-id is not specified, this parameter is required. A backup location that matches this version will be picked to do the restore from. dataset-name-or-id obj-name-or-id
Name or ID of the data set to restore part of. paths dp-path-info[]
List of paths to restore within the data set. Each path can be a single file or a directory. The maximum number of paths is 1000. Output Name Range Type Description job-id integer
Id of the job that handles the restore operation.
Range: [1..2^31-1]
Errno Description EOBJECTNOTFOUND EACCESSDENIED EINTERNALERROR EINVALIDINPUT EDATABASEERROR
| Element definition: dp-path-info | [top] |
Information about a path inside a data set.
Name Range Type Description member-name-or-id obj-name-or-id
Name or id of the data set component that is at the root of the file tree. This may be qtree, ossv directory or volume id that is source of the physical data protection relationship. path string
Name of the path. The maximum path length is 32767 characters. This path is relative to "member-name-or-id" path in the filer or OSSV host. Path cannot be empty. primary-snapshot-name string
optional
Name of the Snapshot copy on the primary where the data being restored originated. This is ignored if primary-snapshot-unique-id is specified. primary-snapshot-unique-id string
optional
Unique id of the Snapshot copy on the primary where the data being restored originated. Currently, this is the Snapshot copy's creation time. This is used to differentiate between versions of the same file in the same backup. If not specified and different versions of a file exist in two Snapshot copies in the backup, then a restore might be ambiguous.
| Element definition: dp-recover-member-info | [top] |
Information about a member inside a data set.
Name Range Type Description destination-host obj-name-or-id
Host to restore this data set member to. The destination host must be compatible with the member type of the member being restored. For example, if the member is a volume, then this destination host must be a filer. destination-path string
Path to restore this data set member to. The path will be interpreted relative to destination-host.
If the destination host is a filer, the path must not already exist. If the destination host is an OSSV, the path will be overwritten.
The format of this value depends on the host type of destination-host. If destination-host is a filer, then it should be of the form "/vol1/[qtree1][/dir1..]". If a full volume restore is being specified, then the destination path must be a volume "/vol". If a full qtree restore is specified, and if the destination is a filer, destination path can be either "/vol" or "/vol/qtree" and if it is "/vol", the server automatically appends the qtree name to the destination path to turn it into "/vol/qtree" format. If destination-host is an OSSV host, then it should be a pathname appropriate for the OSSV host. A Windows host will accept paths such as "C:\My Documents" and a UNIX host will accepts paths such as "/home/user".If the specified dataset-name-or-id is of an application data set, all the above rules apply. In addition, the files will be restored under a uniquely named directory under the destination path. The directory name is created by adding together three separate pieces of information (listed below) separated by a "_" character:
- Name of the primary Snapshot where data originated. All characters except [A-Z], [a-z], [0-9], '.', '-', '_', ')' and '(' are stripped from the Snapshot name before it is used.
- The DFM identifier of the primary volume holding the Snapshot.
- The DFM identifier of the Snapshot. This is currently the timestamp when the Snapshot was created.
The maximum destination length is 32767 characters. In some cases, shorter destination lengths will fail due to limitations of underlying storage systems.member-name-or-id obj-name-or-id
Name or identifier of the data set component that is at the root of the file tree. This may be qtree, ossv directory or volume id that is source of the physical data protection relationship. path string
Name of the path. The maximum path length is 32767 characters. This path is relative to "member-name-or-id" path in the filer or OSSV host. Path cannot be empty. "/" is considered as full member restore when the member is a volume, qtree or OSSV directory. primary-snapshot-name string
optional
Name of the Snapshot copy on the primary where the data being restored originated. This is ignored if primary-snapshot-unique-id is specified. primary-snapshot-unique-id string
optional
Unique id of the Snapshot copy on the primary where the data being restored originated. Currently, this is the Snapshot copy's creation time. This is used to differentiate between versions of the same file in the same backup. If not specified and different versions of a file exist in two Snapshot copies in the backup, then a restore might be ambiguous.
| Element definition: dp-timestamp | [top] |
Seconds since 1/1/1970 in UTC. Range: [0..2^31-1]. This runs out in 2036, so update the API some time before then.
[none]
| Element definition: obj-name-or-id | [top] |
Name or internal ID of a DFM object. This typedef is an alias for the builtin ZAPI type string. An obj-name-or-id must contain between 1 and 64 characters, and must conform to one of the following formats: Elements of type obj-name-or-id are used only as inputs to ZAPIs. The value must match either the name or internal ID of an existing DFM object. The ZAPI must specify the object's DFM object type (e.g. data set, host, DP policy, etc.). Some ZAPIs allow the object to be one of several different types.
- It must have the format of an obj-name, or
- It must be the decimal numeric string form of a positive integer whose value is in the range [1..2^31 - 1].
If the format of an obj-name-or-id input element does not conform, or the value does not match the name or ID of an existing object, then generally the ZAPI documents that it fails with error code EOBJECTNOTFOUND. A ZAPI may return more specific error codes. In such cases, the ZAPI specification must document its behavior.
If a ZAPI can accept a null value (e.g. reference to no object at all) for such an element, then the element is declared optional, and the absence of the input element represents a null value.
[none]